eIQ Ubuntu Installation

[Anonymous123]

Hello, 

I am trying to install the eIQ toolkit on a remote SSH board, the DEBIX Model A Board that has the I.MX 8M Plus EdgeVerse Processor and Ubuntu 20.04. 

I downloaded the eIQ toolkit 1.7.3 Ubuntu installer and installed the toolkit on the remote board.

However, I am wondering about how to access the toolkit's command-line tools, and how to access the python version that has been installed with the eIQ toolkit.

I couldn't find documentation that explains how to interact with the tool from terminal.

Thanks in advance. 

[Harvey021]

Hi @Anonymous123 

You can refer to the document: EIQTKUG-1.7.3.pdf (nxp.com)

Best regards

Harvey

[Anonymous123]

Hi @Harvey021 

Thanks for the support, I'm new to the eIQ toolkit, and I really appreciate your help.

Actually, I'm seeking a better runtime on the DEBIX that has the i.MX 8M Plus processor, I need to deploy a model that should be run fast enough, and I guess the eIQ toolkit could help.

My question is: Do we really need this toolkit on the board itself, like can we convert the model using our regular PCs and then use the converted model (DeepViewRT or tflite) on the board?

What do we actually need on the board so we can run inference with the converted models?

Another important question, is it a must to do the conversion on the board, or can we do it on a Windows machine and then deploy it on the board?

Best regards and thanks a lot

[Harvey021]

Hi 

Actually, we install the toolkit on PC(Both of Windows and Ubuntu), please follow up the User Guide I shared, you'll find answer to your questions. 

Best regards

Harvey

[Anonymous123]

Hi @ Harvey021,

Thanks for assisting me with this topic. 

I read the Guide you shared, and I know how to install and use the toolkit on PCs (either Windows or Ubuntu), but I can't find a precise answer to my question in this Guide.

My question is that: after optimizing/converting the model to either 'tflite' or 'deepviewRT' formats using my PC, I want to deploy the model on the DEBIX Model A board, this board has the i.MX 8M Plus processor. What do I really need on this board to run inference of the converted model?

Do I also need to install the eIQ toolkit on the board? and is it a must to do the conversion on the board itself?

Thanks in advance,

Best regards,